You are on page 1of 56

Pretreatment Role Expectations, Alliance, and Outcome

A dissertation presented to the faculty of the College of Arts and Sciences of Ohio University

In partial fulfillment of the requirements for the degree Doctor of Philosophy

Candace L. Patterson August 2010 © 2010 Candace L. Patterson. All Rights Reserved.

2
 

This dissertation titled Pretreatment Role Expectations, Alliance, and Outcome

by CANDACE L. PATTERSON

has been approved for the Department of Psychology and the College of Arts and Sciences by

________________________________________________________ Timothy M. Anderson Associate Professor of Psychology

________________________________________________________  Benjamin M. Ogles Dean, College of Arts and Sciences

3
 

ABSTRACT PATTERSON, CANDACE L., Ph.D., August 2010, Psychology Pretreatment Role Expectations, Alliance, and Outcome (56 pp.) Director of Dissertation: Timothy M. Anderson Data from 68 clients treated in a naturalistic setting were used to test a meditational model wherein the alliance was hypothesized to mediate the relationship between clients’ pretreatment role expectations and psychotherapy outcome. This proposed model was tested with the procedures developed by Baron and Kenny (1986). All three expectations factors (Personal Commitment, Facilitative Conditions, Counselor Expertise) were related to the alliance. Only expectations for Counselor Expertise related to outcome, although a Sobel test revealed that this relationship was not mediated by the alliance. Suggested research directions, clinical implications, and study limitations are discussed.

Approved:_______________________________________________________________ Timothy M. Anderson Associate Professor of Psychology

4
 

TABLE OF CONTENTS Abstract ................................................................................................................................3 List of Tables .......................................................................................................................5 List of Figures ......................................................................................................................6 Introduction ..........................................................................................................................7 Methods..............................................................................................................................13 Data Analysis .....................................................................................................................20 Results ................................................................................................................................22 Discussion .........................................................................................................................26 References .........................................................................................................................33 Appendix A: Methods .......................................................................................................42 Appendix B: Statistical Analyses ......................................................................................54

5
 

LIST OF TABLES Table 1. Descriptive Statistics and Partial Correlations of EAC-B Factor Scores (Personal Commitment, Facilitative Conditions, Counselor Expertise), WAI-SR Total Score, and OQ Score at Termination, while Controlling for Baseline OQ.........................38 2. Step 1 of Mediation Analyses: Hierarchical Regression Analysis with Counselor Expertise Predicting Outcome while Controlling for Baseline OQ Scores ..............39 3. Step 2 of Mediation Analyses: Hierarchical Regression Analysis with Counselor Expertise Predicting the Alliance while Controlling for Baseline OQ Scores .........40 4. Descriptive Statistics and Partial Correlations of EAC-B Factor Scores (Personal Commitment, Facilitative Conditions, Counselor Expertise), Therapist-Rated WAISR Total Score, Baseline OQ, and OQ Score at Termination ..................................57

..... Mediation of the relationship between clients’ expectations for Counselor Expertise and therapy outcome by the working alliance.. The regression coefficient for this direct path decreased minimally when the indirect path through the alliance was included in the regression equation.......... unmediated relationship)............ *p < ............. All values are standardized regression coefficients (βs)..........41 ....... **p < .05..... after controlling for baseline symptoms.. The value in parentheses is the coefficient for the regression of outcome on Counselor Expertise (the direct.6   LIST OF FIGURES Figure 1.......01 ...

and Outcome Role expectations. Kirsch. Lastly. Counseling expectations were first researched in the 1950s and over the years the amount of research concerning this topic has waxed and waned. 2001). expectations reflect a pantheoretical construct and the implications of research on expectations are likely to be . 1950). refer to a client characteristic that describes one’s anticipatory beliefs about the contributions or behaviors of both the therapist and client in therapy (Nock & Kazdin. Constantino. Following this research and theory is the second major reason why pretreatment expectations warrant further study. 1965. with recently renewed interest reflected in the fact that two review articles on expectancies were published during the past five years (Dew & Bickman.g.7   INTRODUCTION Pretreatment Role Expectations. Farina & Ring. & Bruce. First. The role of treatment expectations in psychotherapy process and outcome warrants further study for three primary reasons. 1999). Kelley. Simply stated. & Ray. Theory suggests that expectations influence one’s perceptions in such a manner that interpersonal encounters and relationships are distorted in order to match one’s expectations (Asch. 1988. 1999). Tinsley. 1946. Therefore. 2006). research indicates that expectations have a powerful influence on individuals’ perceptions and experiences (Greenberg. the therapeutic alliance and clinical outcomes are likely to be shaped by the client’s preexisting expectations about the specific roles of therapist and client during the therapeutic process. Reis & Brown.. Constantino. 2006. Greenberg. also known as counseling or treatment expectations. research indicates that this pretreatment client characteristic can be modified with relative ease (e. 2005. Alliance. Bowman. & Bruce.

(3) the therapist’s empathic understanding and involvement with the client. alliance. Horvath & Symonds. A large body of research demonstrates that clients’ perceptions of the alliance during the early stages of therapy (i. and some research indicates that there is a relationship between clients’ pretreatment role expectations. and (4) the agreement between client and therapist on the tasks and goals for therapy. a modifiable characteristic. 2004. Because it seems likely that role expectations. 2005. (2) the client’s motivation and ability to collaborate with the therapist. Because the alliance is a relational construct and role expectations partially refer to the relational aspects of therapy. including the expected contributions of therapist and client. The published studies . this area of research merits further study (Dew & Bickman. the alliance encompasses both client and therapist contributions to the therapeutic relationship. sessions 3 through 5) consistently predict therapy outcome and account for 7% to 17% of the variance in clinical outcomes (Beutler et al. 1991. and outcome. 2006). & Davis. Constantino. Garske.8   relevant for the majority of clinicians and clients.. Gaston (1990) and Wampold (2001) note that the alliance is comprised of four factors: (1) the emotional aspects of the therapeutic relationship. regardless of the specific intervention being delivered.. Greenberg.e. Thus. Horvath & Bedi. Martin. By gaining an understanding of the specific expectations that impact therapy process and outcome. 2000). influence the alliance and outcome. 2002. researchers and clinicians will be able to investigate and better address the specific role expectations that shape clients’ experiences in therapy as well as their outcomes. it follows that expectancies should be researched in the context of the alliance. & Bruce.

Uhlin. Hardin. genuine. Replicating some of Tokar and colleagues findings. In another study. 1996). Patterson. & Ogrodniczuk. Tokar. & Anderson. Expectations concerning the counselor’s degree of expertise and . 1998. meaning that the more the client expected the therapist to be directive. Uhlin. and help the client solve problems then the lower the clients’ perceived agreement with the clinician on the tasks of therapy. Tokar and colleagues (1996) found that counseling center clients who expected to take responsibility for working in therapy were more likely to view their relationship with the therapist as collaborative and productive. Joyce & Piper. Tokar and colleagues (1996) found that expectations concerning whether the counselor would be accepting. & Brandel. the expectation for counselor expertise was inversely related to the task component of the alliance. McCallum. and trustworthy were not related to client-rated alliance. Adams. 2008. Joyce. Piper. However. offer advice. 2000.9   examining role expectations and the alliance indicate that these constructs do share a relationship with one another and effects sizes for this relationship range from small to large (Al-Darmaki & Kivlighan. and Anderson (2008) found that clients who entered therapy expecting to be responsible for the work involved in treatment were more likely to form a strong working alliance (measured at session 3) with the therapist. Al-Darmaki and Kivlighan (1993) found that university counseling center clients’ relationship expectations (expect to engage in spontaneous self-disclosure while in an egalitarian therapeutic relationship) assessed after three therapy sessions were significantly related to clients’ perceptions of the alliance. Lastly. but these expectations did not relate to therapist-rated alliance. Patterson. 1993.

While findings from these expectations and alliance investigations indicate that expectations and alliance share a relationship. and (4) some studies did not include well-established measures of expectations or alliance (Joyce.. 1993). Tokar. one is limited to drawing tentative conclusions about the specific expectations that impact the alliance due to the following reasons: (1) the disparate types of expectations measured across studies. Joyce. Across studies. Research concerning role expectations and therapy outcome has yielded promising findings. Joyce & Piper. Arnkoff. and Ogrodniczuk (2000) expanded the research on role expectations and the working alliance by including the client’s quality of object relations (QOR) in their study and by examining changes in the alliance across therapy sessions. n = 37. Glass. client improvement. 2000.10   expectations regarding the therapist’s level of genuineness and trustworthiness were not unique predictors of client-rated alliance. For . revealed a significant and positive relationship between expectations and outcome. Piper. & Brandel.. HLM analyses revealed that for high-QOR clients. Hardin. In a review of the expectancy literature. Piper. (2) in one study clients’ expectations were measured after clients attended therapy (Al-Darmaki & Kivlighan. the strength of the relationship between expectations and outcome is characterized by small and medium effect sizes. the lower the pretreatment expectation of being able to contribute to the process of therapy. 1996). continuation in therapy. then the greater the increase in alliance across therapy sessions. McCallum. McCallum. Adams. and Shapiro (2002) found that the majority of the 37 studies examining treatment expectations and outcome (e.g. client satisfaction with therapy. (3) some studies had small sample sizes (e. & Ogrodniczuk. 1998).g. therapeutic alliance).

Results indicate that expectations for client openness and counselor trustworthiness share a positive relationship with therapy satisfaction. 1998). Similar to the research on expectations and the alliance. but this body of research contains certain limitations. & Skinner. social adjustment) in a sample of clients who participated in a controlled trial of short-term individual therapy. these expectations did not predict treatment satisfaction. but when entered in a multiple regression analysis. open-ward psychiatry patients with more negative expectations about receiving guidance. Overall.. In another study. albeit still substantial. and outcome (e. Joyce & Piper. the findings from investigations of expectations and outcome suggest that further research would be valuable due to the fact that there appears to be a relationship between treatment expectations and outcome. symptomatic improvement. . Muller. Joyce and Piper (1998) examined clients’ pretreatment expectations for the usefulness and comfort of a typical therapy session. Anderson. such as the use of measures that are not well-established (e.g. and concern had poorer outcomes (coping patterns. understanding. the alliance. the relationship between clients’ expectations and outcome was smaller than. the relationship between expectations and the alliance. level of functioning) than clients who did not endorse these negative treatment expectations (Jacobs. Heppner and Heesacker (1983) examined the relationships between counseling center clients’ pretreatment role expectations and client satisfaction with therapy.g.11   example. 1972). Results of regression analyses suggest that the expectation that sessions will be productive may help the formation of a productive therapy relationship (expectations accounted 18-40% of the variance in alliance) and expectations contributed approximately 7-10% of the variance in therapy outcome..

Muller. changes in the alliance across sessions (Joyce et al. clients’ perceptions of the alliance during the early phase of therapy.12   In spite of methodological limitations. 1998). 1983). The primary aim of the current study was to advance understanding about the relationships between clients’ pretreatment role expectations. research indicates that role expectations relate to client-rated alliance (AlDarmaki & Kivlighan. Uhlin. 1998). 1998). it may be that client expectations have a direct effect on the working alliance and an indirect effect on clinical outcomes. Tokar et al. & Anderson. alliance. client improvement in terms of behaviors in areas such as impulse control and interpersonal relationships (Jacobs. or predict. and outcome. With the exception of one study by Joyce and Piper (1998).. and outcome provide evidence for the theory that at least some aspects of role expectations relate to. client satisfaction with therapy (Heppner & Heesacker. and clinical outcomes. the relationship between expectations and the alliance was stronger than the relationship between expectations and outcome (Joyce & Piper. 2008. alliance. & Anderson. 1996). 2000). A more specific aim of this study was to apply the Baron and Kenny (1986) approach to test a meditational . To summarize. Patterson. and the relationship between role expectations and outcome in isolation from one another. alliance and outcome. and improved levels of adjustment in social and intimate relationships (Joyce & Piper. Thus. 1972). symptom reduction (Joyce & Piper. 1993. investigators have examined the relationship between role expectations and alliance. and it is possible that the therapeutic alliance mediates the relationship between expectations and outcome.. In the single study that examined expectations. the results from all of the aforementioned studies on expectations.

Clients. the following methodological steps were included in this study: (1) counseling expectations. (2) clients were included in the .4%). sessions 3 through 5) consistently predict treatment outcome. (2) expectations were assessed prior to clients’ first therapy session. This outpatient clinic is housed in the Department of Psychology at a mid-sized university. originally proposed (but not tested) by Dew and Bickman (2005). 64 clients did not meet criteria for participation in the study. wherein the working alliance is hypothesized to mediate the relationship between clients’ pretreatment role expectations and therapy outcomes. and (4) outcome was assessed at the intake and final sessions. further understanding of the role of counseling expectations in therapy process and outcome. This clinic serves both university students (no fee) and community members (small fee). and therapy outcome were assessed with well-established instruments that have adequate psychometric properties. In order to test this model. and advance prior research in this area..13   model. Of the 132 clients that received one or more therapy sessions during the 45month data collection period and consented to having their data used for research purposes.e. METHOD Participants The archival data used in this study came from a research and training clinic that routinely collects data for research and training purposes. Clients were excluded from the study for the following reasons: (1) clients did not complete the EACB prior to the first therapy session. the alliance. (3) alliance was assessed after the third therapy session because early alliance ratings (i. n = 15 (11.

Adjustment Disorder (n = 11). Of the 16 clients that were community members. and they were treated and terminated again) and only data from the first series of therapy sessions were included in this study.8%) did not complete the OQ at baseline. Posttraumatic Stress Disorder (n = 3). Clients attended an average of 12.1%). Clients ranged in age from 18 to 77 and the mean age of clients was 23.8%).5%). Depressive Disorder NOS (n = 1). 10 were employed and 6 were unemployed or retired.14   database more than one time (i.77) therapy sessions.96 (SD = 9. with a range from 4 to 63 sessions..59). they were treated at the clinic. Dysthymic Disorder (n = 6). n = 41 (31.e.85). Anxiety Disorder NOS (n = 1). n = 2 (1. and 8 clients did not provide information regarding their marital status. 26 males). n = 5 (3. terminated. 7 graduate students). and the remaining 7 clients did not provide information about ethnicity. The sample for the present study included 68 clients (42 females.71 (SD = 8. V-Code (n = 11). 3 clients were married. The average baseline OQ score for clients was 71. Most of the clients (n = 52) were university students (13 freshman. Bipolar Disorder (n = 1). African-American (n = 1). 9 sophomores. Bulimia Nervosa (n = 1). which indicates that on average clients entered treatment in the . Hispanic (n = 2). (3) clients did not complete the WAI at session 3. Generalized Anxiety Disorder (n = 8). 3 clients were divorced. and (5) one client (0. The majority of the clients were single (n = 54). Alcohol Abuse (n = 1). and diagnostic data was not available for 7 clients.66 (SD = 24. (4) clients did not attend four or more therapy sessions. 12 seniors. Clients identified as Caucasian (n = 58). The primary diagnoses of clients included the following: Major Depressive Disorder (n = 15). 11 juniors.

with each therapist treating an average of 2. . and 2 therapists were over the age of 40. 3 therapists were in their fourth year of training. and 1 therapist had more than 15 years of clinical experience.30 points. Most clients are referred to the clinic by the university’s counseling center. which indicates that on average clients evidenced reliable change in treatment. Therapists. The number of clients seen by any one therapist ranged from 1 to 6.15   clinical range. The mean change on the OQ from baseline to termination was 15. Typically students in this training clinic received 1-hour of individual and 2-hours of group supervision each week. 3 therapists were between the age of 31 and 40. In terms of level of training. The majority of the clients treated at this clinic present with depressive and anxiety disorders. 4 therapists were in their third training year. 23 therapists were in their second year of clinical training. or instructor and then present to the clinic directly (without a referral). however interpersonal. Of the 33 therapists that participated in this study. 4 were graduate students enrolled in a social work program. Others learn about the clinic through a friend. 2 therapists were in the fifth training year. Information regarding theoretical orientation was not available. and 1 clinician was a clinical psychology professor. psychodynamic. A total of 33 therapists (21 females. 28 were graduate students enrolled in a clinical psychology doctoral program. cognitive-behavioral. Community members typically learn about the clinic through word-of-mouth. academic advisor.51).09 clients (SD = 1. The majority of the therapists (n = 28) were between the age of 20 and 30. 12 males) treated clients that participated in this study. and eclectic theoretical orientations were represented by the supervisors in this clinic.

Facilitative Conditions. Tinsley. feel safe and openly express emotions. 2000. 1982). Hatchett & Han. 2006. clients’ responses to the EAC-B were scored according to the three factor solution of the 53-item EAC-B that was found by Ægisdottier and colleagues (2000). I expect to see an experienced counselor. thus. & Tinsley.. Several factor-analytic studies indicate that a three-factor solution best fits the EAC-B. Ægisdottier. This three factor solution consists of Personal Commitment. Another scoring option is for researchers to score and interpret 53-items from the EAC-B (i. 1989. both the 66-item and 53-item versions (e. Holt. 1991). Tinsley.16   Measures Expectations About Counseling – Brief Form (EAC-B. Gerstein.. Each item is rated on a 7-point fully anchored scale that ranges from 1 (not true) to 7 (definitely true). remain in counseling . Items are prefaced by either “I expect to…” or “I expect the counselor to…” The 66-item EAC-B contains an experimental realism scale that consists of items that reflect the typical practices of the clinic where the measure is being administered. Personal Commitment assesses clients’ expectations for the following aspects of therapy: taking responsibility for personal decisions and the tasks or ‘work’ that is done outside of therapy sessions. & Gridley. and Counselor Expertise. Hayes & Tinsley.e. score all items except those items that comprise the realism scale). the validity of the realism scale is unknown.g. For this study. Some of the items from the realism scale include the following: I expect to take psychological tests. The EACB is a 66-item self-report measure of one’s expectations about counseling. and I expect do assignments outside of the counseling interview. Hinson.

enjoy being with the therapist. Overall. In sum. Counselor Expertise reflects clients’ expectations about whether the therapist will help the client identify and solve problems. the Facilitative Conditions factor measures clients’ expectations for the therapist’s attributes (warm. the Counselor Expertise factor assesses expectations that the . Facilitative Conditions taps clients’ expectations concerning whether the therapist will be a calm. use the counseling relationship to learn how to relate to others and practice new ways of solving problems. and expectations about whether the therapist will self-disclose attitudes and experiences when these attitudes and experiences relate to the client’s difficulties. understand the client’s feelings when the client has not verbally expressed his or her feelings or when the client is unable to translate his or her feelings into words. the Personal Commitment factor corresponds to expectations about being committed and responsible for the work of therapy and expectations about using the relationship with the counselor as practice for relating with individuals outside of the therapy relationship. and improve relationships by gaining a better understanding of people. Taken together. nurturing) and expectations concerning whether the therapist will help in the identification of problems. genuine.17   for at least a few sessions even if it is a painful process. which occasionally requires the therapist to be confrontational. genuine. warm. as well as point out discrepancies between the client’s thoughts and behaviors. and supportive person who will help in the identification of feelings and problematic behaviors. trustworthy.

and help the client identify and solve problems. Tinsley (1982) reported that the internal consistency of the EAC-B scales ranges from 0. with higher scores indicating more distress.92). the EAC-B scales had test-retest reliability ranging from 0. The total score for the OQ ranges from 0 to 180.93).82). 1989. and social role performance (9 items.90). Umphres.87 (empathy scale) with a median test-retest reliability of 0. engage in self-disclosure when relevant.70). Counselor Expertise (α = 0. α = . Response options range from 0 (never) to 4 (almost always). which represents a global measure of functioning. In this study. the internal consistency of the three EAC-B factors was as follows: Personal Commitment (α = 0. Tinsley et al. the total score was included in . A number of studies have focused on the psychometric properties and construct validity of the EAC-B (e. with a median of 0. The Outcome Questionnaire-45 (OQ-45. Lambert. Hansen. These three subscales have adequate internal consistency: subjective discomfort (22 items. This 45-item self-report instrument measures patient progress in therapy. Hatchett & Han. & Burlingame. Concurrent validity has been demonstrated with well-established clinical measures.69 to 0. These studies indicate that the EAC-B has adequate psychometric properties and is a valid measure of expectations about counseling.82. The OQ has three subscales: subjective discomfort. Lunnen..g. interpersonal relations (11 items. has adequate internal consistency (α = .18   therapist will be active and directive. 1982).. and Facilitative Conditions (α = 0.47 (responsibility scale) to 0.74). have insight into the client’s feelings.71 (Tinsley. Hayes & Tinsley. α = . α = . With a two-month interval. 1994). For the present sample. interpersonal relations.91). 2006. 1991). The total score.76. and social role performance.

Items are rated on a 5-point Likert scale. the coefficient alpha for the total score was 0. the WAI-SR total score was utilized in data analyses involving the alliance. Calculations based on these samples indicate that the WAI-SR has adequate internal consistency (α = . The Working Alliance Inventory-Short Form Revised (WAI-SR. This 12-item self-report instrument assesses the strength of the therapeutic alliance.91. Procedure . agreement on goals. The first sample consisted of 231 individuals from a large Midwestern university’s adult psychotherapy clinic.95. and the strength of the bond between client and therapist. For the tasks subscale. These three subscales measure agreement on tasks. The WAI-SR was developed with two different client samples.92.85.87. coefficient alphas are . Coefficient alphas for the goals subscale are .19   analyses involving outcome and the coefficient alpha for the total score in the present sample was 0. In the present sample. Hatcher & Gillaspy. The second sample consisted of 235 persons from a group of counseling centers and outpatient clinics located in the southwest.92) for the total alliance score. A total score and three subscale scores can be derived from the WAI-SR.90 and .85. A total score of 64 or above indicates that it is significantly likely that the individual’s score belongs to a clinical sample and reliable change is set at 14 points. For this study. α = . 2006). Coefficient alphas for the bond subscale are .85 and .87 and .

.. pretreatment role expectations) must be significantly associated with the criterion variable (i. data were collected at an outpatient training clinic previously described under the ‘participants’ heading of this paper. Step 2. therapy outcome). The predictor variable (i.20   Using a naturalistic design. completed the WAI-SR at the conclusion of session 3. The predictor variable (i. the Baron and Kenny (1986) statistical procedure for testing a mediational model was applied to the data.. As part of routine practice. According to Baron and Kenny (1986). Therapists had access to their clients’ data and the clinic maintained a de-identified database of clients’ intake packets and all measures completed at the clinic. testing a mediational model requires the following: Step 1.e. and the WAI-SR immediately after each therapy session. clients completed the EAC-B prior to the intake session. the OQ45 prior to each therapy session. pretreatment role expectations) must be . DATA ANALYSIS Plan of Analysis To examine whether the alliance mediates the relationship between clients’ pretreatment role expectations and outcome.e. and the WAI-SR. Clients were included in the present sample if they completed the EAC-B prior to the first therapy session. the OQ-45. Clients received an intake packet that included a consent form that provided clients with the opportunity to accept or decline participation in archival research projects that utilize the clinic data. completed the OQ-45 at the beginning of the intake and termination sessions. including the EAC-B.e. and attended a total of four or more therapy sessions.

Step 4. There were no missing data for any of the variables included in the analyses and outlier cases were not present in the data. Counselor Expertise). outliers. which is the test of the mediational relationship between the predictor.. Facilitative Conditions. and criterion variable. This difference is assessed with a z test (Preacher & Hayes. alliance). Sobel. and fit between the distribution of the study variables and the assumptions of multiple regression. 2004. missing values. all data were examined with SPSS Version 16 for accuracy of data entry.e. The regression coefficient for the predictor variable in Step 3 must be significantly smaller than the regression coefficient for the predictor variable in Step 1. and the OQ at intake and termination. proposed mediator. These variables included the three expectations factors (Personal Commitment. Step 3. Preliminary analyses revealed that the data did not violate the assumptions of . 1982). A simultaneous regression analysis (criterion variable is regressed on the predictor and the proposed mediator) must reveal a significant relationship between the potential mediator and the criterion variable and the relationship between the predictor and criterion must be smaller than it was in Step 1 (partial mediation) or the relationship between the predictor and the criterion must disappear (complete mediation). the WAI. Assumptions Prior to conducting data analyses.21   significantly associated with the proposed mediating variable (i.

Because baseline OQ was correlated with both the alliance (r (68) = -. and partial correlations (controlling for baseline symptoms) for the EAC-B factors. ranges. but a normal distribution was not obtained. and the WAI were non-normal. and termination OQ are presented in Table 1.60. multicollinearity. the relationships between initial symptoms (baseline OQ). RESULTS Preliminary Analyses In order to determine any differences in expectations between clients included in the study and clients who were excluded from the study. More specifically. Prior to conducting the mediational analyses. p < . or homoscedasticity. standard deviations.21.05). Means. there was not a significant difference for clients included in the study and clients who were excluded from the study on expectations for Personal Commitment (t (130) = -. According to Cohen’s (1992) guidelines.26. p = ns). and termination OQ were examined with Pearson bivariate correlations. WAI-SR total score. Facilitative Conditions (t (130) = -.05) and termination OQ (r (68) = . a series of t-tests were conducted with expectations factors as the dependent variables. correlations between the . Personal Commitment. the alliance. or Counselor Expertise (t (130) = . p = ns). p = ns). Facilitative Conditions. Transformations were applied to these variables. Therefore all of the analyses were performed on the untransformed variables which aid interpretation of the results. baseline OQ was controlled for in all further analyses.22   linearity. Results of these analyses indicate that the expectations of clients who were included in the study did not differ from the expectations of clients who were not included in the study.68. p < .31.

38). analyses for the proposed mediational model with Facilitative Conditions were not possible.32). A large effect size was found for the significant relationship between the alliance and outcome (r = -.60). because expectations for Facilitative Conditions and outcome were not related to one another. termination OQ with the effects of baseline OQ controlled for). Tests for Mediation Step 1: Association Between Predictor (Expectations for Counselor Expertise) and Outcome Results from the regression analyses conducted for Step 1 of the Baron and Kenny (1986) approach to tests of mediation are presented in Table 2.. medium. whereas the relationship between expectations for Counselor Expertise and outcome was significant.e. regression analyses were conducted to determine whether the predictor variable was significantly (and directly) related to outcome. and outcome. Facilitative Conditions. Based on the significant correlations between expectations for Counselor Expertise. analyses for the proposed mediational model with expectations for Personal Commitment were not possible. Expectations for Personal Commitment and expectations for Facilitative Conditions were not related to outcome (i.23   expectations factors (Personal Commitment. the alliance. A medium effect size was found for the relationship between expectations for Counselor Expertise and outcome (r = -. Similarly. For the first step in testing for mediation.27 to . Baseline OQ was . this expectation factor was included in the tests for mediation. Because Personal Commitment and outcome were not related to one another. and large effect sizes (rs ranged from . Counselor Expertise) and the alliance reflect small.

Thus. after controlling for the significant contribution (9. Expectations for Counselor Expertise accounted for a significant amount (6. Counselor Expertise had a direct relationship with the alliance and satisfied the second criterion for mediation. Step 2: Association Between Predictor (Expectations for Counselor Expertise) and the Potential Mediator (Alliance) For the second step in testing for mediation.4%) of the variance in outcome. Step 3: Outcome Regressed on Predictors ((Expectations for Counselor Expertise) and the Potential Mediator (Alliance) Results of the regression analyses for Step 3 of the Baron and Kenny (1986) approach for tests of mediation are presented in Figure 1. The third step in testing for mediation requires that a regression analysis (outcome regressed on expectations and alliance while controlling for baseline OQ) results in a significant relationship between .5%) of baseline OQ scores.5%) of the variance in alliance. Expectations for Counselor Expertise was entered in the second step and explained a significant amount (5. Baseline OQ was entered in the first step of the regression analysis and accounted for a significant 9.5% of the variance in the alliance. These findings indicate that Counselor Expertise had a direct relationship with outcome and thus the first criterion for mediation was met. a regression analysis was conducted to determine whether expectations for Counselor Expertise would predict the alliance when controlling for baseline OQ (see Table 3).24   entered in the first step of the hierarchical regression analysis and accounted for a significant 46% of the variance in outcome. after controlling for baseline OQ.

which suggests complete mediation. indicates whether the difference between the regression coefficient for the relationship between the predictor and outcome is smaller in Step 3 than it was in Step 1.73. 2004. the third criterion for mediation was met for the Counselor Expertise expectations factor. Results of the sobel test were not significant (Sobel z value = -1. Results of a regression analysis with Counselor Expertise and the alliance entered simultaneously as predictors of outcome. Results of this regression analysis also revealed that in the presence of the alliance. Step 4: Decrease in the Regression Coefficients for Expectations The fourth and final criterion for a test of mediation is that the regression coefficient for the predictor (Counselor Expertise) in Step 3 must be significantly smaller than the regression coefficient for the predictor (Counselor Expertise) in Step 1. p = . Sobel.08). More specifically. and that the relationship between expectations and outcome must decrease relative to the strength of the relationship in Step 1 (outcome regressed on expectations). Thus. Thus. 1982). these . the standardized regression coefficients revealed that the strength of the relationship between Counselor Expertise and outcome was smaller than the relationship observed in Step 1. revealed that the alliance was a significant predictor of outcome in the presence of Counselor Expertise. a z test (Preacher & Hayes. the z test determines whether the alliance mediates the relationship between expectations and outcome after controlling for baseline symptoms. expectations for Counselor Expertise did not account for a significant amount of variance in outcome.25   alliance and outcome. This finding suggests that the alliance is not a mediator of the relationship between expectations for Counselor Expertise and outcome (see Figure 1). In other words. This test.

Greenberg. & Bruce. DISCUSSION Despite the knowledge that pretreatment role expectations reflect a malleable client characteristic that appear to relate to both the alliance and treatment outcome. all three types of clients’ pretreatment role expectations were related to the alliance. To my knowledge. research on the role of treatment expectations in therapy process and outcome has been minimal and often limited by methodological and measurement issues (Dew & Bickman. alliance. and clinical outcomes in a naturalistic setting. Although expectations for Counselor Expertise predicted therapy outcome. 2006). Constantino. Only expectations for Counselor Expertise related to treatment outcome. Findings from this study indicate that after controlling for baseline symptoms. this relationship was not explained through the proposed mediating . In order to investigate these relationships. tests of mediation were conducted on data from a sample of 68 clients who attended therapy at an outpatient training clinic that routinely collects client data for research and training purposes. clients’ perceptions of the alliance at the third therapy session. this study represents the first test of mediation for pretreatment role expectations. 2005.26   findings indicate that expectations for Counselor Expertise do not impact treatment outcome via the therapeutic alliance. and therapy outcome. The purpose of this study was to address the limitations found in some of the prior investigations and advance this area of research by examining the relationships between clients’ pretreatment role expectations (expectations for Personal Commitment. Facilitative Conditions. Counselor Expertise).

Patterson. a more recent study found that expectations for Facilitative Conditions and the alliance were correlated with one another (Patterson. further empirical work is needed to determine whether expectations concerning the therapist’s warmth and genuineness relate to the quality and strength of the therapeutic alliance. Although the majority of the research suggests that there is a relationship between these constructs. one prior study found that these constructs did not share a relationship (Tokar et al. & Anderson. a growing body of research suggests that clients’ pretreatment expectations concerning the commitment and work involved in therapy are critical to alliance development. Uhlin. Additionally. Therefore the hypothesis that the alliance . Taken together. genuine. 1996.27   pathway of the therapeutic alliance. the primary study hypothesis (that the alliance would mediate the relationship between expectations and outcome) was not supported. Uhlin. 2008). More specifically. The correlational findings indicated that expectations for Personal Commitment (expect to be committed to and responsible for the work of therapy) and expectations for Facilitative Conditions (expect the therapist to be warm. nurturing) were related to the alliance is partially consistent with results reported in previous research (Tokar et al. with effect sizes for these relationships ranging from medium to large. & Anderson... Research concerning the relationship between expectations for Facilitative Conditions and the alliance has resulted in mixed findings. 1996). these expectations were not related to outcome. whereas another study revealed an association between relationship expectations (expectations that are similar to Facilitative Conditions) and the alliance. Thus. Although expectations for Personal Commitment and expectations for Facilitative Conditions were related to the alliance. 2008).

Martin. & Davis. and helpful in solving problems) was directly related to both alliance and outcome. In sum. However. tests of the proposed mediational model in the present study indicate that clients’ pretreatment role expectations for Counselor Expertise facilitate positive clinical outcomes. In the presence of the significant predictive relationship between alliance and outcome. a Sobel test revealed that the primary study hypothesis that the alliance would mediate the effect of expectations on outcome was not supported. Consistent with findings reported by Joyce and Piper (1998). As was predicted. Garske. results of these analyses support the large body of research (e. clients’ pretreatment expectations for Counselor Expertise (expect the therapist to be directive. the relationship between expectations and the alliance were stronger than the relationship between expectations and therapy outcome. with higher expectations relating to stronger alliances and better outcomes. Also. Taken together. but these effects on outcome are not brought about by the working alliance. expectations for Counselor Expertise no longer predicted outcome..28   would mediate the relationship between expectations and outcome could not be tested with the Personal Commitment and Facilitative Conditions factors. results of the regression analyses that were used to test for mediation revealed that clients who enter therapy with strong expectations that the therapist will be a directive expert are likely to form collaborative and productive bonds with the therapist during the early stage of therapy. Lastly. these findings advance our current understanding of the contributing client characteristics (pretreatment expectations) to therapy process (the . insightful. 2000) that demonstrates that strong alliances contribute to positive clinical outcomes.g.

additional client and therapist variables should be included in future research on the role of expectations in therapy process and outcome. Blatt. alliance. alliance (session 3). it may be that another pretreatment client characteristic. Thus. and outcome (symptoms at baseline and termination).. role expectations represent one of the few client variables (e. 2002) that have been identified as impacting both the alliance and outcome. pretreatment level of perfectionism in the treatment of depression. Although these findings advance prior research. they also leave unanswered questions as to how expectations impact outcome. Several research implications can be drawn from this study’s results. such as attachment style. it could be that clients’ expectations influence a therapist variable such as empathy.g. In addition to determining other factors that may impact expectations. Alternatively. and therapy outcome. and empathy influences both the alliance and outcome. it appears that clients’ pretreatment role expectations reflect a client characteristic and common factor of therapy that influence the relational process of therapy. future . there are implied causal links between expectations (pretreatment). as well as the outcome of therapy. alliance. and progress in therapy. Further conceptual and empirical work is needed to elucidate how expectations contribute to clinical outcome. it is important to note that causality is not certain and other constructs or processes may be responsible for the relationships found among role expectations. However. & Zuroff. Therefore. Shahar. Based on the timing of the assessment of the variables in the present study.29   alliance) and treatment outcome (amelioration of symptoms). is responsible for one’s scores on the expectations factors. alliance ratings. and outcome. Most importantly. For example.

. help the client identify and solve problems). 1999). expert).30   research should examine the most effective methods for shaping clients’ role expectations early in treatment because as was previously mentioned.. Reis & Brown. openly express emotions during therapy. the characteristics of the therapist (e. and the responsibilities of the therapist (e. if the therapist notices difficulties forming a strong alliance early in treatment. role inductions and in-session explorations of expectations should include the following aspects of the roles involved in therapy: the commitment and responsibilities of the client (e. she or he will be more likely to form a strong working alliance with the therapist and enhance the outcome of treatment. work on difficulties outside of therapy). Bowman.. trustworthy. expectations should be addressed by having clients view or listen to a role induction and/or by having therapists initiate a discussion about the expected and actual roles of the client and therapist during treatment. warm. & Ray. the client’s expectations should be attended to during therapy sessions. 1988. help the client identify and label feelings.g. Although further research is needed to determine the best methods for addressing clients’ expectations.g. a key feature of role expectations is the relative ease with which they can be modified and shaped (Tinsley. tentative clinical recommendations are provided. To facilitate a strong alliance. Findings from this study suggest that prior to the initial treatment session or during the first few therapy sessions. Additionally. .g. Once a client’s expectations for the characteristics and responsibilities of the therapist are aligned with the therapist roles described in the preceding sentence.

the use of a single self-report measure of outcome limits the conclusions that can be drawn from this study. More specifically. including symptom distress. The lack of diversity in the participating clinicians and clients limits the generalizability of this study’s results to similar training clinics that treat similar clients. The OQ-45 was designed to measure progress in therapy and captures several domains of client functioning.31   Limitations Although this study addressed limitations found in previous investigations of role expectations.. that the therapist will create facilitative therapeutic conditions (i. Conclusions The strength and quality of the therapeutic relationship relates to clients’ expectations for the roles of both the client and the therapist in treatment. and that the therapist will be knowledgeable and helpful in . The client sample was equally homogeneous and primarily consisted of Caucasian college students. and outcome. Future research could be strengthened by the inclusion of additional outcome measures that tap other domains pertinent to therapy outcome and include multiple raters. clients who expect that they will be committed to and responsible for the work of therapy. warmth. Nevertheless.e. alliance. all of the clinicians were graduate students and the majority of these student-therapists were in their first year of clinical training. Clients’ outcomes were measured with a single self-report instrument (OQ-45). With the exception of 1 therapist. there were limitations of the current study that must be noted. trust. Another important limitation of this study concerns the relatively homogeneous sample of therapists. nurturance). and social role performance. interpersonal functioning.

we will advance our understanding of the numerous variables and mechanisms responsible for the effectiveness of therapy. process. but this relationship is not mediated by the alliance. the relational processes that occur during treatment. and outcome variables in future psychotherapy studies. clients who expect the therapist will knowledgeable and helpful in solving problems will likely have better outcomes than clients who do not have these expectations. By incorporating pretreatment. Additionally.32   solving problems are more likely to form strong and collaborative therapeutic relationships. and the final outcome of therapy. these expectations should be addressed early in treatment to facilitate the process of therapy and the resultant beneficial outcomes. . Because clients’ role expectations reflect a modifiable client characteristic. The findings from this study contribute to our understanding of the relationships between a client characteristic (expectations) that is present prior to the initial therapy session.

R. C. (1993). Beutler. 379-384. & Kenny. A. Journal of Personality and Social Psychology.. L.. Jr. Glass. Noble.).C. L. D. 227-306). J. Forming impressions of personality. Expectations and preferences. S. In. Norcross (Ed. Asch. Inc. Bergin and Garfield’s handbook of psychotherapy and behavior change (pp. Journal of Abnormal and SocialPsychology.. H. & Kivlighan. and statistical considerations. 51. strategic. B. E.. (2000). (2004). Arnkoff. M. Al-Darmaki. M. H. Gerstein. 33. Baron. (1946). R. M. Journal of Counseling Psychology. T. New York: Oxford University Press.33   REFERENCES Ægisdottier. Therapist variables. M. S.335-356). Malik. et al. Lambert (Ed. J. In J. The factorial structure of the Expectations About Counseling Questionnaire – Brief Form: Some serious questions.. 320. (1986). D. New York: John Wiley & Sons.. The moderator-mediator variable distinction in social psychological research: Conceptual. F. & Gridley. D. 40. Talebi. E. Measurement and Evaluation in Counseling and Development. 41. Alimohamed S. 258-290. M. E.. (2002). 1173-1182.).. S. S. B.. Harwood. Congruence in client-counselor expectations for relationship and the working alliance. Psychotherapy relationships that work (pp. & Shapiro.. .

155-159. (2006). Development and evaluation of new factor scales for the Expectations About Counseling inventory in a college sample. Farina.. Greenberg. Anaclitic/sociotropic and introjective/autonomous dimensions. J. A. (1990). A. Are patient expectations still relevant for psychotherapy process and outcome? Clinical Psychology Review. Shahar. (2006). 657-678. Cohen. In J. L.34   Blatt. E. R. & Ring. J. 112.. D. J. (1989). & Zuroff. S. E. 62(10). The influence of perceived mental illness on interpersonal relations. H. G. 36(4).315-334). New York: Oxford University Press.. M. Journal of Clinical Psychology. C. Client expectancies about therapy. G. & Tinsley. 27(2). K. (2006). (2005). R. T. A power primer. . Hayes. P. Hatchett. Gaston. & Gillaspy. (1992). Hatcher. Psychological Bulletin.. Psychotherapy Research. 12-25. 16(1).Identification of the latent dimensions of instruments that measure perceptions of and expectations about counseling. L. 7(1). K. 492-500. Psychotherapy relationships that work (pp. Development and validation of a revised short version of the working alliance inventory (WAI-SR). A. J. Constantino. 26. 70. 143-153. (2002). Psychotherapy. C. Norcross (Ed. The concept of the alliance and its role in psychotherapy: Theoretical and empirical considerations. S. Journal of Counseling Psychology. 1303-1318. L. T. Journal of Abnormal Psychology. Dew. J. (1965).). & Bruce. Mental Health Services Research. 21-33. N. 47-51. & Bickman. & Han.

31-39. Journal of Counseling Psychology. (1972). 139-149. J. Psychotherapy relationships that work (pp. W. New York: Oxford University Press.37-69). 30(1). (1999). Journal of Personality. Kirsch. Journal of Psychotherapy Practice and Research. S. Relation between working alliance and outcome in psychotherapy: A meta-analysis.. Piper. P. 39(3). (1950). E. Joyce. W. Kelley. 431-439. Journal of Psychotherapy Practice and Research. A. Journal of Counseling and Clinical Psychology. In J. D. H. 18. O. premorbid adjustment. 7. 213-225. O. (2002). H. M. S. . Therapeutic expectations. S. and treatment outcome in short-term individual psychotherapy. 9. the therapeutic alliance. Muller. (1998). 236-247.. & Skinner. J. A.).35   Heppner. Horvath. Warm-cold variable in first impressions of persons. C. & Bedi. and manifest distress level as predictors of improvement in hospitalized patients. P.. Norcross (Ed. Role behavior expectancies and alliance change in short-term individual psychotherapy. & Piper. & Ogrodniczuk. Expectancy. B.. Perceived counselor characteristics. 455-461. and client satisfaction with counseling. Washington. Jacobs. (1983). The alliance. McCallum.. J. (Ed. A. & Heesacker. J. R. How expectancies shape experience. J. client expectations. Horvath. A. A. M. & Symonds.). 38(2). M. Journal of Counseling Psychology. I. Joyce.. P.. E. DC: American Psychological Association. C. (2000). Anderson. (1991).

Sociological methodology (pp. F. M. Reducing psychotherapy dropouts: Maximizing perspective convergence in the psychotherapy dyad. M. T. M. Patterson. V.. UT: IHC Center for Behavioral Healthcare Efficacy. Administration and Scoring Manual of the Outcomes Questionnaire (OQ-45. N. L. Psychotherapy.. Nock. Behavior Research Methods. Asymptomatic confidence intervals for indirect effects in structural equation models. K. 10. P.. Hansen. Clients’ pretreatment counseling expectations as predictors of the working alliance. & Anderson. L. Journal of Child and Family Studies. Umphres. 68(3). SPSS and SAS procedures for estimating indirect effects in simple mediation models. 55(4). & Kazdin. Instruments. Reis.). 36(4). Preacher. In S. Leinhart (Ed. G.36   Lambert. & Brown. (2004). 36(2). (2001). 123136. . K. M. 438-450. B. & Burlingame. E. J. (2000). Salt Lake City. J. & Computers. G.. Journal of Counseling Psychology. 155-180. K. & Davis. F. (1994). Garske. 717-731. (1982). K..1). 528-534. (2008). San Francisco: Jossey-Bass. Journal of Consulting and Clinical Psychology. 290-312). C. A.. Lunnen. Uhlin... A. J. Martin. Parent expectancies for child therapy: Assessment and relation to participation in treatment. & Hayes. B. E. D. Relation of the therapeutic alliance with outcome and other variables: A meta-analytic review. (1999). Sobel.

A. Department of Psychology. 99108. E. M. Southern Illinois University at Carbondale. B. & Tinsley. J. W. E. I. & Brandel. E.. E. J. B. M.. Tinsley. & Ray. 101-110. S. 9-26. H. D. Wampold.. Tinsley. (1991). 24(3). A. Hinson.. A. Manipulation of expectancies about counseling and psychotherapy: Review and analysis of expectancy manipulation strategies and results. Holt. E. Expectations about counseling. I. Unpublished test manual. The great psychotherapy debate. Clients’ expectations about counseling and perceptions of the working alliance. (1988). L. S. . H. D. Journal of Counseling Psychology... 35(1). (1996). A. A construct validation study of the Expectations About Counseling-Brief Form: Factorial validity. 11. (2001). London: Lawrence Erlbaum Associates.. M. Tokar. Hardin. Measurement and Evaluation in Counseling and Development. H. (1982).37   Tinsley.. Bowman. Journal of College Student Therapy. S. Adams. S.

00-133.66 0. Counselor Expertise 4.37 1.35 26.00 0.38   Table 1 Descriptive Statistics and Partial Correlations of EAC-B Factor Scores (Personal Commitment.01.44-7. while Controlling for Baseline OQ Variable 1.82 5. **p < . 1 ___ 2 3 4 5 0.19 0.05.00 0.17 . and OQ Score at Termination.49** -0. N = 68. Termination OQ M SD Range Note. WAI-SR 5.94 2.06 2.44** 0. WAI-SR Total Score. Personal Commitment 2.00 ___ -0.10-7. Facilitative Conditions.22 5.32* 4.72** ___ 0.73** ___ 0.05 1. Facilitative Conditions 3.  *p < .27* -0.60 2.38** 4.67-7.60** -0.19 1. Counselor Expertise).00 ___ 56.18 5.33-5.

740 0.46 56. 65 0. *p < .099 0. 66 .16* 1.235 7.678 7.51 34.303 0.68* 0.639 -0.29** -2.096 1.697 -5.981 0.46 56.Table 2 Step 1 of Mediation Analyses: Hierarchical Regression Analysis with Expectations Factors Predicting Outcome while Controlling for Baseline OQ Scores Model Variable B SE B β t R2 F Step ΔR2 ΔF df Outcome (Termination OQ) Step 1 Baseline OQ Step 2 Baseline OQ Counselor Expertise Note.25** 0.05. **p < .25** 1. 0.33** 0. N = 68.05 7.50** 0.01.

N = 68.308 -2.003 0.30* 2.17** 0.058 -0.265 0.095 6.93* 1.23* -0.16 6.003 -0.40   Table 3 Step 2 of Mediation Analyses: Hierarchical Regression Analysis with Expectations Factors Predicting the Alliance while Controlling for Baseline OQ Scores Model Variable B SE B β t Alliance Step 1 Baseline OQ Step 2 Baseline OQ Counselor Expertise Note.008 0.63* 0.131 0. 65 0. **p < -0. *p < .07 4.10 6.258 -2. 66 R2 F Step ΔR2 ΔF df .05.93* 0.99* 1.006 0.

*p < .41   Alliance   .26* -. unmediated relationship). after controlling for baseline symptoms.24)* -. Mediation of the relationship between clients’ expectations for Counselor Expertise and therapy outcome by the working alliance. **p < .01.29** Expectations for Counselor Expertise (-.05. The regression coefficient for this direct path decreased minimally when the indirect path through the alliance was included in the regression equation. All values are standardized regression coefficients (βs). The value in parentheses is the coefficient for the regression of outcome on Counselor Expertise (the direct. ns Outcome Figure 1.17. .

Appendix A: Methods .

43   .

44   .

45     .

46     .

47     .

48     .

49     .

50     .

51     .

52     .

53                                       Appendix B: Statistical Analyses .

expectations. 1991) was calculated for the current sample. The majority (66. It should be noted that 5 (26. Results indicate that 56. Results indicate that therapists’ 3rd session alliance ratings do not share a relationship with baseline OQ.3%) of these clients began therapy in the non-clinical range.8%) of the sample met criteria for clinically significant change. the relationships between initial symptoms (baseline OQ). or termination OQ (see Table 1). A total of 19 (27.1% (n = 2) of clients evidenced reliable deterioration. and termination OQ were examined with Pearson bivariate correlations. Approximately 40% (n = 26) of clients showed reliable improvement and 3. . pretreatment role expectations. as opposed to a non-clinical population.9% (n = 37) of the clients did not evidence reliable change. clients had to show reliable change on the OQ and fall in the functional range (OQ score falling between 0 and 63 points) at termination. Exploratory Analyses In order to understand the role of therapist-rated alliance. therapist-rated alliance.54   Reliable Change and Clinical Significance An OQ score of 64 or above indicates that the respondent’s score is more likely to come from a clinical. To be classified as having made clinically significant change. Jacobson & Truax. the Reliable Change Index (RCI. reliable change is reflected in a change (in either direction) of 14 OQ points from baseline to termination.2%) of clients in the present study began therapy with OQ scores reflecting a clinical population. For the OQ. To evaluate whether client change on the OQ was reliable.

00 0.01.00 ___ 0.04 1.94 0.025* -0. Counselor Expertise 4.05.33** 4. Counselor Expertise).65** 71.00135. Baseline OQ.02 -0.08 2.37 1.55   Table 4 Descriptive Statistics and Partial Correlations of EAC-B Factor Scores (Personal Commitment. Therapist-Rated WAI-SR Total Score.69 0.73** ___ .48** 0. N = 65.  *p < .00 ___ 55. Facilitative Conditions 3.34** 5. Personal Commitment 2.50 2.00 Note.12 -0.16 23.99 10.12 3. and OQ Score at Termination Variable 1. Facilitative Conditions.21 1.107.14 -.09 -0.675. Baseline OQ 6. Termination OQ M SD Range 0.77** ___ 0.447. 1 ___ 2 3 4 5 6 0.27* -0.31* 5.99 25.17 .677.95 2.92 5.00 ___ -0.15 -0. WAI-SR 5. **p < .00133.

. N. Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. P. S. & Truax. Journal of Consulting and Clinical Psychology. 59(1). (1991). 12-19..56   Reference Jacobson.